When enabling swap space, it's important to ensure the volume's IO performance is high to maintain acceptable overall performance.

Recommendations for AWS

Use Instance Store instead of EBS volume to mount the swap data. For more details, refer to the Instance Store swap volumes documentation.

Implications of swap usage

If your workload exceeds the physical memory and begins to use swap, the JCC node can slow down due to the additional IO overhead. To mitigate this, configure higher timeouts for jcc.status_timeout_seconds and jcc.jcc_poll_timeout_seconds for health checks.

Resource exhaustion

Even with swap configured, if all available memory is used up, the JCC process may still run out of resources. This can cause the JCC process to restart, terminating all running pipelines. To prevent this, use larger nodes with more memory to ensure your workload can complete successfully.

Swap configuration

Limit the swap size to the maximum that the JCC node will use. Setting a larger swap size can degrade performance, particularly during Java Runtime Environment (JRE) garbage collection operations.

• Memory swapping can result in performance degradation because of disk IO, especially if the Pipeline workload also utilizes local disk IO.

• When the pipeline workload is dynamic and capacity planning is difficult, we strongly recommend the minimum swap configuration.

To utilize swap for the JCC process, you can use this workflow:

1. Enable swap on your host machine. The steps depend on your operating system (OS). For example, for Ubuntu Linux, you can use the steps in this tutorial.

2. Update your Maximum memory Snaplex setting to a lower percentage value, such that the absolute value is lower than the available memory. The load balancer uses this value when allocating Pipelines. The default is to set to 85%, which means that if the node memory usage is above 85% of the maximum heap size, then additional pipelines cannot start on the node.

If a node is swapping, execution times might get slower over time. You can use these two properties to avoid the monitor from terminating the JCC process because of health check timeout. While not specific to swapping, they might help increase when swap is used. In the Global properties section of the Node Properties tab in the Snaplex page:

• jcc.jcc_poll_timeout_seconds is the timeout (default value is 10 seconds) for each health check poll request from the Monitor.

• jcc.status_timeout_seconds is the time period (default value is 300 seconds) that the Monitor process waits for before the JCC is restarted if the health check requests continuously fail.